summaryrefslogtreecommitdiffstats
path: root/doc/Hash.3
blob: f9f4a394444a6e452eb3aea28604c2f3d745bfdb (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
'\"
'\" Copyright (c) 1989-1993 The Regents of the University of California.
'\" Copyright (c) 1994-1996 Sun Microsystems, Inc.
'\"
'\" See the file "license.terms" for information on usage and redistribution
'\" of this file, and for a DISCLAIMER OF ALL WARRANTIES.
'\" 
'\" RCS: @(#) $Id: Hash.3,v 1.5 2000/07/19 22:15:28 ericm Exp $
'\" 
.so man.macros
.TH Tcl_Hash 3 "" Tcl "Tcl Library Procedures"
.BS
.SH NAME
Tcl_InitHashTable, Tcl_InitHashTableEx, Tcl_InitObjHashTable, Tcl_DeleteHashTable, Tcl_CreateHashEntry, Tcl_DeleteHashEntry, Tcl_FindHashEntry, Tcl_GetHashValue, Tcl_SetHashValue, Tcl_GetHashKey, Tcl_FirstHashEntry, Tcl_NextHashEntry, Tcl_HashStats \- procedures to manage hash tables
.SH SYNOPSIS
.nf
\fB#include <tcl.h>\fR
.sp
\fBTcl_InitHashTable\fR(\fItablePtr, keyType\fR)
.sp
\fBTcl_InitHashTableEx\fR(\fItablePtr, keyType, typePtr\fR)
.sp
\fBTcl_InitObjHashTable\fR(\fItablePtr\fR)
.sp
\fBTcl_DeleteHashTable\fR(\fItablePtr\fR)
.sp
Tcl_HashEntry *
\fBTcl_CreateHashEntry\fR(\fItablePtr, key, newPtr\fR)
.sp
\fBTcl_DeleteHashEntry\fR(\fIentryPtr\fR)
.sp
Tcl_HashEntry *
\fBTcl_FindHashEntry\fR(\fItablePtr, key\fR)
.sp
ClientData
\fBTcl_GetHashValue\fR(\fIentryPtr\fR)
.sp
\fBTcl_SetHashValue\fR(\fIentryPtr, value\fR)
.sp
char *
\fBTcl_GetHashKey\fR(\fItablePtr, entryPtr\fR)
.sp
Tcl_HashEntry *
\fBTcl_FirstHashEntry\fR(\fItablePtr, searchPtr\fR)
.sp
Tcl_HashEntry *
\fBTcl_NextHashEntry\fR(\fIsearchPtr\fR)
.sp
char *
\fBTcl_HashStats\fR(\fItablePtr\fR)
.SH ARGUMENTS
.AS Tcl_HashSearch *searchPtr
.AP Tcl_HashTable *tablePtr in
Address of hash table structure (for all procedures but
\fBTcl_InitHashTable\fR, this must have been initialized by
previous call to \fBTcl_InitHashTable\fR).
.AP int keyType in
Kind of keys to use for new hash table.  Must be either
TCL_STRING_KEYS, TCL_ONE_WORD_KEYS, TCL_CUSTOM_TYPE_KEYS,
TCL_CUSTOM_PTR_KEYS, or an integer value greater than 1.
.AP Tcl_HashKeyType *typePtr in
Address of structure which defines the behaviour of the hash table.
.AP char *key in
Key to use for probe into table.  Exact form depends on
\fIkeyType\fR used to create table.
.AP int *newPtr out
The word at \fI*newPtr\fR is set to 1 if a new entry was created
and 0 if there was already an entry for \fIkey\fR.
.AP Tcl_HashEntry *entryPtr in
Pointer to hash table entry.
.AP ClientData value in
New value to assign to hash table entry.  Need not have type
ClientData, but must fit in same space as ClientData.
.AP Tcl_HashSearch *searchPtr in
Pointer to record to use to keep track of progress in enumerating
all the entries in a hash table.
.BE

.SH DESCRIPTION
.PP
A hash table consists of zero or more entries, each consisting of
a key and a value.
Given the key for an entry, the hashing routines can very quickly
locate the entry, and hence its value.
There may be at most one entry in a hash table with a
particular key, but many entries may have the same value.
Keys can take one of three forms:  strings,
one-word values, or integer arrays.
All of the keys in a given table have the same form, which is
specified when the table is initialized.
.PP
The value of a hash table entry can be anything that fits in
the same space as a ``char *'' pointer.
Values for hash table entries are managed entirely by clients,
not by the hash module itself.
Typically each entry's value is a pointer to a data structure
managed by client code.
.PP
Hash tables grow gracefully as the number of entries increases,
so that there are always less than three entries per hash bucket,
on average.
This allows for fast lookups regardless of the number of entries
in a table.
.PP
\fBTcl_InitHashTable\fR calls the extended function 
\fBTcl_InitHashTableEx\fR with a NULL \fItypePtr\fR.
.PP
\fBTcl_InitHashTableEx\fR initializes a structure that describes
a new hash table.
The space for the structure is provided by the caller, not by
the hash module.
The value of \fIkeyType\fR indicates what kinds of keys will
be used for all entries in the table.  \fIKeyType\fR must have
one of the following values:
.IP \fBTCL_STRING_KEYS\fR 25
Keys are null-terminated ASCII strings.
They are passed to hashing routines using the address of the
first character of the string.
.IP \fBTCL_ONE_WORD_KEYS\fR 25
Keys are single-word values;  they are passed to hashing routines
and stored in hash table entries as ``char *'' values.
The pointer value is the key;  it need not (and usually doesn't)
actually point to a string.
.IP \fBTCL_CUSTOM_TYPE_KEYS\fR 25
Keys are of arbitrary type, and are stored in the entry. Hashing
and comparison is determined by \fItypePtr\fR. The Tcl_HashKeyType 
structure is described in the section 
\fBTHE TCL_HASHKEYTYPE STRUCTURE\fR below.

.IP \fBTCL_CUSTOM_TYPE_KEYS\fR 25
Keys are pointers to arbitrary type, and the are stored in the entry. Hashing
and comparison is determined by \fItypePtr\fR. The Tcl_HashKeyType 
structure is described in the section 
\fBTHE TCL_HASHKEYTYPE STRUCTURE\fR below.
.IP \fIother\fR 25
If \fIkeyType\fR is not one of the above,
then it must be an integer value greater than 1.
In this case the keys will be arrays of ``int'' values, where
\fIkeyType\fR gives the number of ints in each key.
This allows structures to be used as keys.
All keys must have the same size.
Array keys are passed into hashing functions using the address
of the first int in the array.
.PP
\fBTcl_InitObjHashTable\fR uses \fBTcl_InitHashTableEx\fR to
initialize a hash table whose keys are Tcl_Obj *.
.PP
\fBTcl_DeleteHashTable\fR deletes all of the entries in a hash
table and frees up the memory associated with the table's
bucket array and entries.
It does not free the actual table structure (pointed to
by \fItablePtr\fR), since that memory is assumed to be managed
by the client.
\fBTcl_DeleteHashTable\fR also does not free or otherwise
manipulate the values of the hash table entries.
If the entry values point to dynamically-allocated memory, then
it is the client's responsibility to free these structures
before deleting the table.
.PP
\fBTcl_CreateHashEntry\fR locates the entry corresponding to a
particular key, creating a new entry in the table if there
wasn't already one with the given key.
If an entry already existed with the given key then \fI*newPtr\fR
is set to zero.
If a new entry was created, then \fI*newPtr\fR is set to a non-zero
value and the value of the new entry will be set to zero.
The return value from \fBTcl_CreateHashEntry\fR is a pointer to
the entry, which may be used to retrieve and modify the entry's
value or to delete the entry from the table.
.PP
\fBTcl_DeleteHashEntry\fR will remove an existing entry from a
table.
The memory associated with the entry itself will be freed, but
the client is responsible for any cleanup associated with the
entry's value, such as freeing a structure that it points to.
.PP
\fBTcl_FindHashEntry\fR is similar to \fBTcl_CreateHashEntry\fR
except that it doesn't create a new entry if the key doesn't exist;
instead, it returns NULL as result.
.PP
\fBTcl_GetHashValue\fR and \fBTcl_SetHashValue\fR are used to
read and write an entry's value, respectively.
Values are stored and retrieved as type ``ClientData'', which is
large enough to hold a pointer value.  On almost all machines this is
large enough to hold an integer value too.
.PP
\fBTcl_GetHashKey\fR returns the key for a given hash table entry,
either as a pointer to a string, a one-word (``char *'') key, or
as a pointer to the first word of an array of integers, depending
on the \fIkeyType\fR used to create a hash table.
In all cases \fBTcl_GetHashKey\fR returns a result with type
``char *''.
When the key is a string or array, the result of \fBTcl_GetHashKey\fR
points to information in the table entry;  this information will
remain valid until the entry is deleted or its table is deleted.
.PP
\fBTcl_FirstHashEntry\fR and \fBTcl_NextHashEntry\fR may be used
to scan all of the entries in a hash table.
A structure of type ``Tcl_HashSearch'', provided by the client,
is used to keep track of progress through the table.
\fBTcl_FirstHashEntry\fR initializes the search record and
returns the first entry in the table (or NULL if the table is
empty).
Each subsequent call to \fBTcl_NextHashEntry\fR returns the
next entry in the table or
NULL if the end of the table has been reached.
A call to \fBTcl_FirstHashEntry\fR followed by calls to
\fBTcl_NextHashEntry\fR will return each of the entries in
the table exactly once, in an arbitrary order.
It is unadvisable to modify the structure of the table, e.g.
by creating or deleting entries, while the search is in
progress.
.PP
\fBTcl_HashStats\fR returns a dynamically-allocated string with
overall information about a hash table, such as the number of
entries it contains, the number of buckets in its hash array,
and the utilization of the buckets.
It is the caller's responsibility to free the result string
by passing it to \fBckfree\fR.
.PP
The header file \fBtcl.h\fR defines the actual data structures
used to implement hash tables.
This is necessary so that clients can allocate Tcl_HashTable
structures and so that macros can be used to read and write
the values of entries.
However, users of the hashing routines should never refer directly
to any of the fields of any of the hash-related data structures;
use the procedures and macros defined here.

.SH "THE TCL_HASHKEYTYPE STRUCTURE"
.PP
Extension writers can define new hash key types by defining four
procedure, initializing a Tcl_HashKeyType structure to describe
the type, and calling \fBTcl_InitHashTableEx\fR.
The \fBTcl_HashKeyType\fR structure is defined as follows:
.CS
typedef struct Tcl_HashKeyType {
    int \fIversion\fR;
    int \fIflags\fR;
    Tcl_HashKeyProc *\fIhashKeyProc\fR;
    Tcl_CompareHashKeysProc *\fIcompareKeysProc\fR;
    Tcl_AllocHashEntryProc *\fIallocEntryProc\fR;
    Tcl_FreeHashEntryProc *\fIfreeEntryProc\fR;
} Tcl_HashKeyType;
.CE
.PP
The \fIversion\fR member is the version of the table. If this
structure is extended in future then the version can be used
to distinguish between different structures. It should be set
to \fBTCL_HASH_KEY_TYPE_VERSION\fR.
.PP
The \fIflags\fR member is one or more of the following OR'ed together:
.IP \fBTCL_HASH_KEY_RANDOMIZE_HASH\fR 25
There are some things, pointers for example which don't hash well 
because they do not use the lower bits. If this flag is set then the
hash table will attempt to rectify this by randomising the bits and 
then using the upper N bits as the index into the table.
.PP
The \fIhashKeyProc\fR member contains the address of a function 
called to calculate a hash value for the key.
.CS
typedef unsigned int (Tcl_HashKeyProc) (
    Tcl_HashTable *\fItablePtr\fR,
    VOID *\fIkeyPtr\fR);
.CE
If this is NULL then \fIkeyPtr\fR is used and 
\fBTCL_HASH_KEY_RANDOMIZE_HASH\fR is assumed.
.PP
The \fIcompareKeysProc\fR member contains the address of a function 
called to compare two keys.
.CS
typedef int (Tcl_CompareHashKeysProc) (VOID *\fIkeyPtr\fR,
    Tcl_HashEntry *\fIhPtr\fR);
.CE
If this is NULL then the \fIkeyPtr\fR pointers are compared.
If the keys don't match then the function returns 0, otherwise
it returns 1.
.PP
The \fIallocEntryProc\fR member contains the address of a function 
called to allocate space for an entry and initialise the key.
.CS
typedef Tcl_HashEntry *(Tcl_AllocHashEntryProc) (
    Tcl_HashTable *\fItablePtr\fR, VOID *\fIkeyPtr\fR);
.CE
If this is NULL then Tcl_Alloc is used to allocate enough space for a
Tcl_HashEntry and the key pointer is assigned to key.oneWordValue.
String keys and array keys use this function to allocate enough 
space for the entry and the key in one block, rather than doing
it in two blocks. This saves space for a pointer to the key from
the entry and another memory allocation. Tcl_Obj * keys use this 
function to allocate enough space for an entry and increment the 
reference count on the object.
If 
.PP
The \fIfreeEntryProc\fR member contains the address of a function 
called to free space for an entry.
.CS
typedef void (Tcl_FreeHashEntryProc) (Tcl_HashEntry *\fIhPtr\fR);
.CE
If this is NULL then Tcl_Free is used to free the space for the 
entry. Tcl_Obj * keys use this function to decrement the
reference count on the object.
.SH KEYWORDS
hash table, key, lookup, search, value